Gemma 4 Chat — Browser LLM

Context8,192 tokens

Modelgemma-4-E2B-it · q4f16

RuntimeWebGPU

PrivacyIn-browser inference

Initializing...

Gemma 4 WebGPU

A multimodal AI assistant running in your browser on your own GPU. Model weights download on first use.

onnx-community/gemma-4-E2B-it-ONNX · q4f16 quantization

+ Image

Max tokens:

Voice off

Press the mic — models load on first use, then just talk. Talk over Gemma to interrupt.

Mic 2x

You'll see your words appear here as you speak…